Aitken's delta-squared process

From Wikipedia, the free encyclopedia
Jump to: navigation, search

In numerical analysis, Aitken's delta-squared process is a series acceleration method, used for accelerating the rate of convergence of a sequence. It is named after Alexander Aitken, who introduced this method in 1926.[1] Its early form was known to Seki Kōwa (end of 17th century) and was found for rectification of the circle, i.e. the calculation of π. It is most useful for accelerating the convergence of a sequence that is converging linearly.

Definition[edit]

Given a sequence x={(x_n)}_{n\in\N}, one associates to this sequence the new sequence

A(x)={\left(\frac{x_{n+2}\,x_n-(x_{n+1})^2}{x_{n+2}-2\,x_{n+1}+x_n}\right)}_{n\in\Z^*},

which can, with improved numerical stability, also be written as

 A(x_n)=x_n-\frac{(\Delta x_n)^2}{\Delta^2 x_n},

where

\Delta x_{n}={(x_{n+1}-x_{n})},\

and

\Delta^2 x_n=x_n -2x_{n+1} + x_{n+2},\

for n = 0, 1, 2, 3, \dots \,

(To use this nice operator notation, one has to allow for the indices to start from n = 2 on, or apply a translation operator which first shifts the sequence indices by two, or to adopt the convention that xn = 0 for all n < 0.)

Obviously, Ax is ill-defined if Δ2x contains a zero element, or equivalently, if the sequence of first differences has a repeating term. From a theoretical point of view, assuming that this occurs only for a finite number of indices, one could easily agree to consider the sequence Ax restricted to indices n>n0 with a sufficiently large n0. From a practical point of view, one does in general rather consider only the first few terms of the sequence, which usually provide the needed precision. Moreover, when numerically computing the sequence, one has to take care to stop the computation when rounding errors become too important in the denominator, where the Δ² operation may cancel too many significant digits.

Properties[edit]

Aitken's delta-squared process is a method of acceleration of convergence, and a particular case of a nonlinear sequence transformation.

x will converge linearly to \ell if there exists a number μ ∈ (0, 1) such that

 \lim_{n\to \infty} \frac{|x_{n+1}-\ell|}{|x_n-\ell|} = \mu.

Aitken's method will accelerate the sequence x_n if \lim_{n \to \infty}\frac{A_n-\ell}{x_n-\ell}=0.

A is not a linear operator, but a constant term drops out, viz: A[x-\ell] = Ax - \ell, if \ell is a constant. This is clear from the expression of Ax in terms of the finite difference operator \Delta.

Although the new process does not in general converge quadratically, it can be shown that for a fixed point process, that is, for an iterated function sequence x_{n+1}=f(x_n) for some function f, converging to a fixed point, the convergence is quadratic. In this case, the technique is known as Steffensen's method.

Empirically, the A-operation eliminates the "most important error term". One can check this by considering a sequence of the form x_n=\ell+a^n+b^n, where 0<b<a<1: The sequence Ax will then go to the limit like b^n goes to zero.

One can also show that if x goes to its limit \ell at a rate strictly greater than 1, Ax does not have a better rate of convergence. (In practice, one rarely has e.g. quadratic convergence which would mean over 30 resp. 100 correct decimal places after 5 resp. 7 iterations (starting with 1 correct digit); usually no acceleration is needed in that case.)

In practice, Ax converges much faster to the limit than x does, as demonstrated by the example calculations below. Usually, it is much cheaper to calculate Ax (involving only calculation of differences, one multiplication and one division) than to calculate many more terms of the sequence x. Care must be taken, however, to avoid introducing errors due to insufficient precision when calculating the differences in the numerator and denominator of the expression.

Example calculations[edit]

The value of \sqrt{2} \approx 1.4142136 can be approximated by assuming an initial value for a_n and iterating the following:
a_{n+1} = \frac{a_n + \frac{2}{a_n}}{2}.

Starting with a_0 = 1:

n x = iterated value Ax
0 1 1.4285714
1 1.5 1.4141414
2 1.4166667 1.4142136
3 1.4142157 --
4 1.4142136 --
The value of \frac{\pi}{4} may be calculated as an infinite sum:
\frac{\pi}{4} = \sum_{n=0}^\infty \frac{(-1)^n}{2n+1} \approx 0.785398
n term x = partial sum Ax
0 1 1 0.79166667
1 −0.33333333 0.66666667 0.78333333
2 0.2 0.86666667 0.78630952
3 −0.14285714 0.72380952 0.78492063
4 0.11111111 0.83492063 0.78567821
5 −9.0909091×10−2 0.74401154 0.78522034
6 7.6923077×10−2 0.82093462 0.78551795
7 -6.6666667×10−2 0.75426795 --
8 5.8823529×10−2 0.81309148 --

Example Pseudocode For Aitken Extrapolation[edit]

The following is an example of using the Aitken extrapolation to help find the limit of the sequence x_{n+1} = f(x_n) when given x_0, which we assume to be the fixed point  \alpha = f(\alpha). For instance, we could have x_{n+1} = \frac{1}{2}(x_n + \frac{2}{x_n}) with x_0 = 1 which has the fixed point \sqrt{2} so that f(x) = \frac{1}{2}(x + \frac{2}{x}) (see Methods of computing square roots).

This pseudo code also computes the Aitken approximation to f'(\alpha). The Aitken extrapolates will be denoted by aitkenX. We must check if during the computation of the extrapolate the denominator becomes too small, which could happen if we already have a large amount of accuracy, since otherwise a large amount of error could be introduced. We denote this small number by epsilon.

%These choices depend on the problem being solved
x0 = 1                      %The initial value
f(x) = (1/2)*(x + 2/x)      %The function that find the next element in the sequence
tolerance = 10^-10          %10 digit accuracy is desired
epsilon = 10^-16            %Don't want to divide by a number smaller than this
 
maxIterations = 20          %Don't allow the iterations to continue indefinitely
haveWeFoundSolution = false %Were we able to find the solution to the desired tolerance? not yet.
 
for i = 1 : maxIterations 
    x1 = f(x0)
    x2 = f(x1)
 
    lambda = absoluteValue((x2 - x1)/(x1 - x0));      %OPTIONAL: computes an approximation of |f'(fixedPoint)|, which is denoted by lambda
 
    denominator = x2 - 2*x1 + x0
 
    if(absoluteValue(denominator) < epsilon)          %Don't want to divide by too small of a number
        print('WARNING: denominator is too small')
        break;                                        %Leave the loop
    end
 
    aitkenX = x2 - ( (x2 - x1)^2 )/denominator
 
    if(absoluteValue(aitkenX - x2) < tolerance)       %If the result is within tolerance
        print("The fixed point is", aitkenX))         %Display the result of the Aitken extrapolation
        haveWeFoundSolution = true
        break;                                        %Done, so leave the loop
    end
 
    x0 = aitkenX                                      %Update x0 to start again                  
 
end
 
if(haveWeFoundSolution == false)   %If we weren't able to find a solution to within the desired tolerance
    print("Warning: Not able to find solution to within the desired tolerance of ", tolerance);
    print("The last computed extrapolate was ", aitkenX)
end

See also[edit]

Notes[edit]

  1. ^ Alexander Aitken, "On Bernoulli's numerical solution of algebraic equations", Proceedings of the Royal Society of Edinburgh (1926) 46 pp. 289–305.

References[edit]